About the method

A perceived Conversational GenAI quality framework

AIPI Wiki is a reflection and research-engagement tool. It asks users to evaluate how a specific conversational GenAI application is perceived across system behavior, information output, visible reasoning or explanation, and provider support.

System Model

Contains 5 perceived dimensions: S1, S2, S3, S4, S5.

Information Model

Contains 6 perceived dimensions: I1, I2, I3, I4, I5, I6.

Reasoning Model

Contains 4 perceived dimensions: R1, R2, R3, R4.

Provider Model

Contains 4 perceived dimensions: P1, P2, P3, P4.

Scoring

1. Each item is answered on a 1-7 agreement scale.
2. The mean score is computed for each of the 19 dimensions.
3. Each dimension becomes L, M, or H.
4. The 19-level vector is compared with predefined templates using Manhattan distance.
5. If the best similarity is below 60%, MIXD is shown with the top three nearest profiles.

Scope

The result is not a clinical or psychological diagnosis. It is a structured way to discuss your perception of a GenAI application in a particular context.

Wording rule: the site describes perceived profiles and user impressions. It does not claim that AI systems possess real personalities.